CommunicationSegregation-Induced Enhancement ofLow-Temperature Tensile Ductility in aCast High-Nitrogen AusteniticStainless Steel Exhibiting Deformation-Induced a¢ Martensite Formation

JAVAD MOLA, MARCO WENDLER, ANDREASWEIß, BENEDIKT REICHEL, GOTTHARD WOLF,and BRUNO C. DE COOMAN

In spite of the formation of a high fraction of defor-mation-induced a¢ martensite, the tensile elongation of acast high-nitrogen austenitic stainless steel was found toenhance at lower temperatures, a behavior deviatingfrom that exhibited by wrought and homogenizedaustenitic stainless steels. The observed behavior wasexplained by the presence of microstructural regionswith different stabilities with respect to deformation-induced a¢ martensite formation caused by the segrega-tion of alloying elements.

DOI: 10.1007/s11661-015-2782-y� The Minerals, Metals & Materials Society and ASMInternational 2015

Tensile elongation near room temperature of lowstacking fault energy (SFE) austenitic steels includingstainless,[1–9] high Mn,[10–13] and high Ni[14] steels varies inthree temperature regimes as follows. At the highesttemperature range (regime I), tensile elongation remainsmore or less constant or exhibits a weak temperaturedependence. At intermediate temperatures (regime II),elongation increases as the temperature decreases. Theenhancement of ductility at reduced temperatures inregime II is commonly attributed to such deformation-induced microstructural changes as e martensite forma-tion, the e-TRIP effect,[15] and deformation twinning, theTWIP effect.[2] These are common microstructural fea-tures of deformed austenitic stainless steels[16,17] and highMn[12,18,19] steels and are thought to be consequences ofhigh glide planarity. Enhanced glide planarity caused by

reduced cross slip of screw dislocations at lower temper-atures has been correlated with an underlying temperaturedependence of SFE.[20] The latter dependence is availablefor many fcc metals and alloys including austeniticsteels.[20–24] The enhancement of tensile elongation atlower temperatures in regime II of austenitic stainlesssteels is interrupted at the temperature below which thedeformation-induced formation of a¢ martensite is en-abled, namely below the Md

cfia¢ temperature. The loss ofductility caused by the deformation-induced formation ofa¢ martensite initiates the regime III of elongation,characterized by reduced elongations at lower tempera-tures. The large number of investigations in support of thedetrimental effect of a¢ martensite formation on the tensileelongation of metastable austenitic stainless steels[1–7, 10–12]

indicates that any possible contribution to the ductilityof a TRIP effect would not be large enough to preventthe loss of ductility below the Md

cfia¢ temperature. Thishas been also suggested by the modeling of the a¢ TRIPeffect contribution to the ductility.[25] The present studydemonstrates how compositional inhomogeneities in acast high-nitrogen austenitic stainless steel can causedeviation from the behavior described above.The chemical composition of the experimental steel

was Fe-14.3Cr-5.5Mn-5.5Ni-0.5Si-0.37N-0.02C(mass pct). The high nitrogen content of the steel wasachieved by melting under nitrogen pressures up to 4bar. The high superheating of the molten alloy (2023 Kor 1750 �C) and the raised temperature of the mold keptinside the pressurized melting chamber resulted in a slowcooling rate during solidification. Tensile specimens witha gage diameter of 6 mm were subsequently machinedfrom the as-cast ingot with a dimension of2109 809 45 mm3. To ensure the absence of deforma-tion-induced martensite at the surface of machinedtensile specimens, they were heat treated at 1323 K(1050 �C) for 30 minutes. Using an initial strain rate of49 10 �4 s �1, tensile specimens were tensile tested attemperatures between 233 K and 473 K (�40 �C and200 �C). According to the prior experience of theauthors, this strain rate is sufficiently low to ensurenegligible adiabatic heating of tensile specimens. Micro-structural examination of tensile-tested specimens wasperformed by means of light optical microscopy, scan-ning electron microscopy (SEM) including electronbackscatter diffraction (EBSD), and transmission elec-tron microscopy (TEM). Quantification of deformation-induced a¢ martensite in the gage area of tensile-testedspecimens was done by magnetic saturation measure-ments using a Metis MSAT magnetic saturation deviceequipped with a Lakeshore 480 fluxmeter. Details on thequantification procedure may be found in.[26]

The optical micrograph in Figure 1(a) shows themicrostructure prior to tensile experiments. InFigure 1(b),the weakly etched grain boundaries in Figure 1(a) aredemarcated by dashed lines. The microstructure consistsof dendrites commonly extending across grain boundaries.Thermodynamic calculations using Thermo-Calc with aTCFE7 database (not shown) predicted an austeniticsolidification for the alloy. Austenitic mode of solidifica-tion was also predicted based on empirical relationshipstaking into account the ratio of ferrite to austenite

JAVAD MOLA, MARCO WENDLER, Scientific Assistants, andANDREAS WEIß, Professor, are with the Institute of Iron and SteelTechnology, Technische Universitat Bergakademie Freiberg, LeipzigerSt. 34, 09599 Freiberg, Germany. Contact e-mail: javad.mola@iest.tu-freiberg.de BENEDIKT REICHEL, Scientific Assistant, is with theInstitute of Materials Science, Technische Universitat BergakademieFreiberg, 09599 Freiberg, Germany. GOTTHARD WOLF, Professorand Director, is with the Foundry Institute, Technische UniversitatBergakademie Freiberg. BRUNO C. DE COOMAN, Professor andDirector, is with the Materials Design Laboratory, Graduate Instituteof Ferrous Technology, Pohang University of Science and Technol-ogy, Pohang 790-784, South Korea.

Manuscript submitted November 12, 2014.Article published online February 10, 2015

1450—VOLUME 46A, APRIL 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

stabilizing alloying elements.[27,28] The dendritic morphol-ogy inFigures 1((a), (b)) therefore reveals the arrangementof primary dendrites of austenite and interdendriticregions formed at later stages of solidification. The darkspots distributed mainly in interdendritic regions inFigures 1((a), (b)), and shown at a higher magnificationin the SEM micrograph of a deformed specimen inFigure 1(c), are primarily undissolved M2N nitrides (Mdenoting Cr, Fe, and substitutional alloying elements).Precipitates have coarsened significantly during the slowcooling of the ingot and the subsequent annealing at1323 K (1050 �C). M2N precipitates persisted in spite oftheir being predicted by Thermo-Calc to fully dissolve at1323 K (1050 �C).

Engineering stress–strain curves for tensile specimenstested at various temperatures are shown in Figure 2(a).At all tensile test temperatures studied, decreasingtemperature was associated with simultaneous improve-ment of strength and ductility. The temperature depen-dences of the 0.2 pct proof stress (Rp0.2), tensile strength(Rm), uniform elongation (Ag), and total elongation (A)are shown in Figure 2(b).

In Figure 2(b), the temperature dependence of defor-mation-induced a¢ martensite fraction (fa¢) determinedby magnetic saturation measurements is superimposedon tensile properties. Deformation-induced a¢ martens-ite only occurred at and below room temperature. Thehighest deformation-induced a¢ martensite content of53 vol pct occurred during tensile deformation at 233 K(�40 �C).

EBSD analysis of selected deformation conditionswas performed to study the occurrence of commondeformation-induced processes such as twinning andmartensite formation. EBSD orientation maps of spec-imens strained at 373 K and 293 K (100 �C and 20 �C)are shown in Figures 3((a), (b)). No twinning or emartensite was detected even after EBSD scans using arefined step size of 0.2 lm. Formation of coarsedeformation twins in Fe-16Cr-6Mn-(3,6,9)Ni alloys,[29]

Fe-19Cr-4Ni-3Mn-0.25C-0.15N alloys,[8,9] and Fe-17Cr-6Mn-0.45C-9Ni-(0,4)Al alloys[30] tensile tested at tem-peratures above their respective Md

cfia¢ temperatureshas been readily confirmed with EBSD scans using stepsizes of the order of 0.2 lm. The absence of deformationtwins is in agreement with the reported twin-suppressingrole of N in Fe-Cr-Ni stainless steels.[31] As shown inFigures 3((c), (d)), 15 pct tensile straining at 233 K(�40 �C) was associated with the formation of both eand a¢ martensites. These phases were always observedin the vicinity of one another. In agreement with themagnetic saturation measurement results, tensile strain-ing until fracture at 233 K (�40 �C) was associated withsignificant a¢ martensite formation (Figures 3((e), (f)).Figure 3(f) which shows only the orientation map of theaustenite phase clearly indicates that transformedregions reproduce the dendritic morphology of Fig-ures 1((a), (b)). The arrows in Figures 3((e), (f)) denote atransformed dendrite core on both sides of an austenitegrain boundary. The deformation-induced transforma-tion of dendrite core regions is attributed to thesegregation of alloying elements during the formationof primary dendrites of austenite. Segregation of major

alloying elements including Cr, Ni, and Mo to thesurrounding liquid phase during the formation ofprimary dendrites of austenite has been observed in anFe-19Cr-14Ni-2.5Mo cast stainless steel.[32] Manganesehas been found to exhibit a similar behavior.[27] Fur-thermore, the higher concentration of undissolvednitrides in interdendritic regions suggests that the

: Undissolved precipitates

Austenitic matrix

20 μm

Austenite grain boundaries

100 μm

(b)

(c)

Dendrite cores 100 μm

(a)

Interdendritic regions

Fig. 1—(a, b) Light optical and (c) scanning electron micrographs ofthe cast steel. Specimens were electrolytically etched with oxalic acid.The contrast in (c) was generated by secondary electrons.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, APRIL 2015—1451

nitrogen too has segregated out of the growing dendritesof primary austenite. Due to the high diffusivity ofnitrogen at the solution annealing temperature of1323 K (1050 �C) which enables its long range diffusion

and homogenization in the austenite, any possibleinhomogeneity of N concentration is thought to onlyarise due to local variations in the solubility of N causedby the segregation of substitutional alloying elements.The distribution of alloying elements had made dendritecore regions favorable sites for the deformation-inducedformation of a¢ martensite. The alloy can therefore be

(a)

(b)

(c)

Fig. 2—(a) Engineering stress–strain curves for the experimental steelat various temperatures; (b) temperature dependence of tensile prop-erties and deformation-induced a¢ content; (c) temperature depen-dence of tensile elongation for Fe-(17-18)Cr-Ni alloys with variousNi concentrations adapted from.[33]

Tensile direction

111

101001

(a)

(e) (f)

(c)

(d)

(b)

γ (green)ε (yellow)

α′ (red)

γ + ε + α′ γ only

100 μm

50 μm

10 μm

50 μm

10 μm

50 μm

IPF373 K(100 °C)39% T.S.

IPF293 K(20 °C)15% T.S.

IPF233 K(-40 °C)15% T.S.

IPF233 K(-40 °C)65% T.S.

Phase233 K(-40 °C)15% T.S.

0001

ε α and γ

Fig. 3—(a through c, e, f) EBSD inverse pole figure (IPF) maps oftensile specimens deformed to indicated tensile strains (T.S.) at tem-peratures marked. Colors indicate crystal directions parallel to ten-sile direction; (d) phase map corresponding to the IPF in (c).

1452—VOLUME 46A, APRIL 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

regarded as being composed of dendrite core regionswith a low stability with respect to the deformation-induced a¢ martensite formation and farther regionswith a higher stability. In other words, the Md

cfia¢

temperature gradually decreases from cores of dendritestoward interdendritic regions.

The localized transformation of dendrite core regionsis thought to be responsible for the observation ofimproved tensile ductility at subzero temperatures wheredeformation-induced a¢ martensite formation is enabled(Figure 2(b)). This behavior is inconsistent with thewell-documented elongation regimes discussed in theintroduction and exemplified with wrought Fe-17Cr-(9-20)Ni stainless steels in Figure 2(c).[33] The markedMd

cfia¢ temperature of Fe-17Cr-(9-20)Ni alloys almostexactly coincides with the temperature of highest tensileelongation. The formation of a¢ martensite below theMd

cfia¢ temperature thus leads to the onset of ductilityregime III characterized by lower elongations at reducedtensile test temperatures.

The temperature dependence of tensile elongation inthe present high N alloy is the balance of that atdifferent regions of the microstructure. In fact, the alloycan be visualized as being composed of regions withdifferent temperature dependencies of tensile elongationwhose summation according to the rule of mixturesgives the overall temperature dependency of tensileelongation depicted in Figure 2(b). On this basis, thetensile test temperature of 293 K (20 �C) should lie inductility regime III of the transformed regions (neardendrite cores), whereas it is still in ductility regime II ofthe more heavily alloyed interdendritic regions. There-fore, although reducing tensile test temperature wouldhave a negative effect on the ductility of dendrite coreregions due to the earlier occurrence of deformation-induced a¢ martensite, it enhances ductility of the moreheavily alloyed interdendritic regions by promotingplanar glide in the absence of deformation-induced a¢martensite formation. The balance of these opposingeffects is thought to be responsible for the observedenhancement of ductility in spite of a¢ martensiteformation at 263 K and 233 K (�10 �C and �40 �C).

Two families of steels which are prone to composi-tional inhomogeneities within austenitic regions areTRIP-assisted and quenched and partitioned (Q&P)steels. Although differences in the stability of austenitein such steels may also arise from differences in suchparameters as the size and morphology of retainedaustenite regions,[34] a scenario analogous to that of thepresent steel with significant local variations in thestability of retained austenite with respect to martensitictransformation can most likely arise due to localvariations in the carbon concentration of retainedaustenite. Accordingly, in studies involving tensile testsat various temperatures of TRIP-assisted and Q&Psteels,[35,36] the observation of highest tensile elongationsat temperatures associated with transformation ofnoticeable fractions of original c content may beattributed to compositional inhomogeneities withinretained austenite rather than a TRIP effect.

The occurrence of e martensite during low-tempera-ture deformation of the alloy, as suggested by the EBSD

phase map of Figure 3(d), was confirmed by comple-mentary TEM examinations. The e martensite platesappear to be closely connected with the a¢ martensiteformed during tensile tests at and below RT. Theselected area diffraction (SAD) pattern in Figure 4(a)whose solution is provided in Figure 4(b) was takenfrom the encircled region in the bright field image ofFigure 4(c) and pertains to a specimen strained 15 pct at233 K (�40 �C). Shown in Figures 4((d) through (f)) aredark field images corresponding to the bright field image

BF

Z.A.γ

ε

α′

BFDF1

DF2

DF3

DF1, ε+α′

DF2, ε DF3, α′

(a) (b)

(c) (d)

(e) (f)

0.2 μm 0.2 μm

0.2 μm0.2 μm

(g) (h)BF DF, ε

α′

α′

ε + γ

0.1 μm 0.1 μm

Fig. 4—(a) SAD pattern of a specimen tensile strained 15 pct at233 K (�40 �C). As shown in (b), the pattern was found to consistof superimposed reflections of c, e, and a¢ phases. The area used forSAD pattern acquisition is encircled in (c). (d through f) are darkfield images recorded using the reflections marked in (a); (g) brightfield image of a specimen deformed until fracture (65 pct T.S.) at233 K (�40 �C); (h) dark field image of the area shown in (g)recorded using a reflection of e phase.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, APRIL 2015—1453

of Figure 4(c). They were recorded using the reflectionsmarked in Figure 4(a). The wide band illuminated in thedark field image of Figure 4(e) was identified as emartensite. The thinner plates oriented more verticallywith respect to the wide band of e martensite which areseparately shown in the dark field image of Figure 4(f)were found to be a¢ martensite. The dark field image inFigure 4(d) was recorded using a common reflection of eand a¢ martensite phases. Although thin plates of a¢martensite can be seen in the bright field image ofFigure 4(c) to extend outside the wide e martensite plate,they do not diffract according to a¢ martensite outsidethe region of intersection. Provided that thinner platesoutside the intersection region are another variant of emartensite, the crystallographic description of deforma-tion-induced cfia¢ transformation via intermediate emartensite[37] may be used to visualize the shearsrequired to accomplish the transformation. TEM exam-ination of a specimen tensile tested until fracture at233 K (�40 �C) revealed the presence of a high densityof e martensite plates in untransformed austeniticregions. This behavior is exemplified in the bright fieldimage of Figure 4(g). The corresponding dark fieldimage in Figure 4(h) was recorded using a reflection of emartensite. The dominant occurrence of e martensite inthe deformed microstructure of the investigated high Nstainless steel suggests that it acts as precursor to thedeformation-induced formation of a¢ at larger strains.

In summary, the temperature dependence of tensileelongation in a high-nitrogen cast stainless steel wascorrelated with microstructural changes during tensiledeformation. The observation of enhanced tensile elon-gation at low temperatures where a¢ martensite formationnear dendrite core regions was enabled was justified bythe presence of interdendritic regions exhibiting superiorductility at reduced tensile test temperatures. The differ-ence in the stability of austenitic regions was caused byelemental segregation during the slow primary solidifica-tion of the alloy. Furthermore, the alloy exhibited a highresistance against deformation twinning. In addition to a¢martensite, deformed microstructures at low deformationtemperatures contained a high density of e martensite.

The support of German Research Foundation inthe framework of Collaborative Research Center 799is gratefully acknowledged.

REFERENCES1. M. Pozuelo, J.E. Wittig, J.A. Jimenez, and G. Frommeyer: Metall.

Mater. Trans. A, 2009, vol. 40A, pp. 1826–34.

2. A.S. Hamada, L.P. Karjalainen, R.D.K. Misra, and J. Talonen:Mater. Sci. Eng. A, 2013, vol. 559, pp. 336–44.

3. A. Jahn, A. Kovalev, A. Weiß, S. Wolf, L. Kruger, and P.R.Scheller: Steel Res. Int., 2011, vol. 82, pp. 39–44.

4. A. Kovalev, A. Jahn, A. Weiß, and P.R. Scheller: Steel Res. Int.,2011, vol. 82, pp. 45–50.

5. A. Kovalev, A. Jahn, A. Weiß, S. Wolf, and P.R. Scheller: SteelRes. Int., 2011, vol. 82, pp. 1101–07.

6. A. Kovalev, A. Jahn, A. Weiß, S. Wolf, and P.R. Scheller: SteelRes. Int., 2012, vol. 83, pp. 576–83.

7. M. Wendler, A. Weiß, L. Kruger, J. Mola, A. Franke, A. Kovalev,and S. Wolf: Adv. Eng. Mater., 2013, vol. 15, pp. 558–65.

8. M. Wendler, J. Mola, B. Reichel, L. Kruger, and A. Weiß: inHNS2014, Hamburg, 2014.

9. M. Wendler, J. Mola, B. Reichel, L. Kruger, and A. Weiß: inHMnS 2014, Aachen, 2014, pp. 407–10.

10. A.S. Hamada, L.P. Karjalainen, and M.C. Somani: Mater. Sci.Eng. A, 2007, vol. 467, pp. 114–24.

11. S. Curtze and V.-T. Kuokkala: Acta Mater., 2010, vol. 58,pp. 5129–41.

12. O. Grassel, L. Kruger, G. Frommeyer, and L.W. Meyer: Int.J. Plast., 2000, vol. 16, pp. 1391–409.

13. M. Hauser: Diploma Thesis, TU Bergakademie Freiberg, 2013.14. I. Tamura: Met. Sci., 1982, vol. 16, pp. 245–53.15. W. Zhang and J. Hu: Mater. Charact., 2013, vol. 79, pp. 37–42.16. L. Remy and A. Pineau: Metall. Trans., 1974, vol. 5, pp. 963–

65.17. H. Biermann, J. Solarek, and A. Weidner: Steel Res. Int., 2012,

vol. 83, pp. 512–20.18. S. Allain, J.-P. Chateau, O. Bouaziz, S. Migot, and N. Guelton:

Mater. Sci. Eng. A, 2004, vols. 387–389, pp. 158–62.19. O. Bouaziz, S. Allain, C.P. Scott, P. Cugy, and D. Barbier: Curr.

Opin. Solid State Mater. Sci., 2011, vol. 15, pp. 141–68.20. F. Abrassart: Metall. Trans., 1973, vol. 4, pp. 2205–16.21. L. Remy, A. Pineau, and B. Thomas: Mater. Sci. Eng., 1978, vol.

36, pp. 47–63.22. R.M. Latanision and A.W. Ruff: Metall. Trans., 1971, vol. 2,

pp. 505–09.23. F. Lecroisey and B. Thomas: Phys. Status Solidi A, 1970, vol. 2,

pp. K217–20.24. L. Remy: Acta Metall., 1977, vol. 25, pp. 173–79.25. H.K.D.H. Bhadeshia: ISIJ Int., 2002, vol. 42, pp. 1059–60.26. M. Wendler, B. Reichel, R. Eckner, O. Fabrichnaya, L. Kruger, A.

Weiß, and J. Mola: Metall. Mater. Trans. A, 2014, DOI: 10.1007/s11661-014-2716-0.

27. P.L. Ferrandini, C.T. Rios, A.T. Dutra, M.A. Jaime, P.R. Mei,and R. Caram: Mater. Sci. Eng. A, 2006, vols. 435–436, pp. 139–44.

28. J.A. Brooks and A.W. Thompson: Int. Mater. Rev., 1991, vol. 36,pp. 16–44.

29. A. Vinogradov, A. Lazarev, M. Linderov, A. Weidner, and H.Biermann: Acta Mater., 2013, vol. 61, pp. 2434–49.

30. R. Rahimi, B.C. De Cooman, H. Biermann, and J. Mola: Mater.Sci. Eng. A, 2014, vol. 618, pp. 46–55.

31. M. Yoshitake, T. Tsuchiyama, and S. Takaki: Tetsu-to-Hagane,2012, vol. 98, pp. 223–28.

32. N. Lewis, M.J. Cieslak, and W.F. Savage: J. Mater. Sci., 1987, vol.22, pp. 2799–2810.

33. H. Gutte and A. Weiß: Habilitation, TU Bergakademie Freiberg,2011.

34. L. Samek, E. De Moor, J. Penning, and B.C. De Cooman: Metall.Mater. Trans. A, 2006, vol. 37A, pp. 109–24.

35. Y. Sakuma, D.K. Matlock, and G. Krauss:Metall. Trans. A, 1992,vol. 23A, pp. 1233–41.

36. W. Feng, Z. Wu, L. Wang, and J.G. Speer: Steel Res. Int., 2013,vol. 84, pp. 246–52.

37. T.H. Lee, H.Y. Ha, J.Y. Kang, J. Moon, C.H. Lee, and S.J. Park:Acta Mater., 2013, vol. 61, pp. 7399–7410.

1454—VOLUME 46A, APRIL 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A